What Actually Works on Site
Centimetre-accurate Augmented Reality, digital twins and construction AI that crews actually use. Why sometimes field software fails and what makes it work outdoors. Augmented reality was supposed to change construction a decade ago. Somewhere at the bottom of that hype cycle, a small group of companies quietly worked out where the technology actually pays off and it turned out to be outside the building, in the trench, next to the utilities nobody can see. Alec Pestov is the founder and CEO of vGIS, the company behind vSite, software that lets field crews see underground infrastructure with centimetre-level accuracy. Alec came from strategy consulting, not utilities, and started the company by his own admission the wrong way round: building interesting technology first and searching for a problem afterwards. This conversation traces what he learned correcting that mistake.
0:00
Andrea Ciaffi
Hi, Alec. Welcome to the Digital Construction Podcast. In this episode, we sit down with Alec Pestov. He's the founder and CEO of VGIS, the company behind VSite, software that helps utility crews see undergrounds. Pretty cool stuff, guys. And it basically makes sure that people keep the field projects on track.
0:25
Andrea Ciaffi
We will be tracing this eight-year journey from the peak of the AR hype cycle down into the trench, where finally the technology started paying off. We see so many very often the new technology have a lot of hype around it, but then eventually we found a way to use this technology on real site.
0:45
Andrea Ciaffi
And Alex company, Alex company is a, it's a good evidence of this type of situation. But before we start, Alec, I have a nice breaker for you. What is the best piece of advice you've ever received in your career?
1:01
Alec Pestov
Don't start your own company.
1:04
Andrea Ciaffi
Yeah.
1:04
Alec Pestov
And this is the, and first of all, thank you for having me today, Andrea. Happy to be on your podcast and happy to share the story. So feel free to ask any question you want. But the best piece of advice, yes, going back to your question is, don't start your own business.
1:18
Alec Pestov
And that's the advice that I'm going to give to anybody and everybody.
1:23
Andrea Ciaffi
I want to ask you more. Why is that?
1:26
Alec Pestov
Well, it's like getting punched in the gut every single day. Unless you enjoy the journey and you're okay with getting kicked in the guts every day, then it's a very unrewarding often place to be.
1:45
Andrea Ciaffi
Yeah, it's a common story. It's very hard. But speaking of that, let's talk about how you found this company. Tell us, take us back to the moment where you first decided to start VAGIS. What convinced you to build this company and what did you not yet understand about the market?
2:04
Andrea Ciaffi
that you are entering.
2:07
Alec Pestov
started it by mistake and started absolutely the wrong way you can start the company. I had another startup, so this is my 4th startup. I had two service-based startups, which did all right. So it didn't make me a billionaire, but gave me the starting capital for my next startup, which failed.
2:24
Alec Pestov
So it was a knowledge management system and it failed because we underestimated the human nature. So the problem with knowledge management is that you need to invest time up front to save time later. And if you can invest 5 minutes today as a human to save 2 hours tomorrow, you will never invest 5 minutes today.
2:47
Alec Pestov
You'll just, you rather struggle for the weekend, like through the weekend or many hours tomorrow. then spend extra 5 minutes today. And the knowledge management didn't take off. And at that time, I was at the crossroads. Others say, I give up and I don't want to do my businesses anymore and go back and work for somebody else or do something with the technology we developed.
3:12
Alec Pestov
And my technical team had expertise some expertise in augmented reality, virtual reality, as we try to adopt our knowledge management platform to the virtual world. And I took the team at the same time, HoloLens came out and I thought, we must do something.
3:27
Alec Pestov
This is new iPhone. It's going to take over the world. And we're going to be like one of those first companies that started the mobile revolution. except on wearable devices. So I have to do it, I have to do it, and with the remaining funds, I bought HoloLens for myself, developers, so funded development, and we started building a product.
3:50
Alec Pestov
So that was the product. We didn't have any domain expertise, so we found somebody who was looking for the... technical team to build what they wanted and combining 2 ideas together, we came up with the first version of our product. So absolutely wrong way, so we were not solving the problem.
4:06
Alec Pestov
We were... developing cool technology hoping that it will solve a problem once it's developed and it's the way I would say I would never start another company again.
4:18
Andrea Ciaffi
Start to the problem first, right?
4:21
Alec Pestov
Yeah, and then search for the solution. Well, search with the solution and search for the problem. That's what we did. And you have to do it other way. Understand the domain, understand the problem, understand lack of suitable solutions,
4:30
Andrea Ciaffi
Mm.
4:33
Alec Pestov
and then offer something that nobody else offers.
4:36
Andrea Ciaffi
Yeah, and you touched on this briefly. You don't come specifically from this industry. Your background is in strategy and consulting, not utility, right? So this is pipes and conduits below ground. How did you close the gap, the industry knowledge gap in the early years?
4:51
Andrea Ciaffi
I'm sure now you're an expert in this field, but how did you manage that there at the beginning of your journey?
5:00
Alec Pestov
Just trying to understand what people wanted. So at the beginning, we didn't target construction space. The first group of people that approached us was from municipal infrastructure projects type of stuff. So they do, they did maintenance of sewer and water infrastructure for municipality.
5:21
Alec Pestov
And I had to understand some of their problems. And working with them, I very quickly understood that there is no future for our product in municipal world. At least not in the form that it had before. There are quite a few demographics problems, procurement cycle problems, So there are multiple issues with targeting that specific industry.
5:45
Alec Pestov
But because of the publicity we received initially, and because of the hype that was around HoloLens and Mr. at the time, construction companies approached us and brought in some ideas. So we extended our tools to cover some of that stuff. And step by step, step by step, we started understanding what we can do.
6:07
Alec Pestov
So I didn't start as an expert. I'm not an expert in it by any means, but we understand our space quite well at this point.
6:17
Andrea Ciaffi
I guess at some point you started receiving some signals that something was a bit different. What was the signal that told you infrastructure construction was different from every earlier lead that you tried before?
6:33
Alec Pestov
We tried everything. So we said the solution as the struggling startup, you have the technology, you try to find the space for that. We tried oil and gas. We even went to hydraulic fracking to visualize drilling in real time through HoloLens. We've done custom work from municipalities.
6:52
Alec Pestov
We've done work for United States Air Force. So we did a variety of different things. And I was always looking for a situation where solution solves a problem so that I can say, I can see how it solves the problem, where it's not just a cool thing to have.
7:11
Alec Pestov
But I can see your problem. I can see how this solution can be the only solution to your problem or the best solution to your problem. And that, of course, made us bounce between different ideas, different fields. So we always stayed in generally the same category, but we experimented in different areas as well.
7:29
Alec Pestov
We tried vertical construction for a little bit without abandoning what we had in the field. But we always went back to our roots as of, okay, we seem to be doing better outside of the building. We seem to be doing better outside of oil and gas.
7:45
Alec Pestov
How do we expand it? What else can we try here? to make it work. And that's a story. So we tried many things, went down the dead ends many, many, many times just to come back to the same original idea, keep developing it.
8:01
Andrea Ciaffi
Yeah, definitely. You said that, don't start your company. It's been a hard journey, I guess. Let's talk about, let's talk about your main product, VSite. What, what, what tell us what is VSite for and what it does?
8:18
Alec Pestov
We currently do two things. So if you simplify what we do, it's a comprehensive platform with many features. But if you distill it down to what we do is we help organize data so it can be used in the field.
8:34
Alec Pestov
And then once people in the field start working with the data, we give them all the tools to document what they do so that it can preserve evidence, and produce necessary documentation for typical infrastructure project. So those are the two things we do, and we're expanding into a third one, where we will use AI and other tools to leverage outputs of the documentation we produce.
9:03
Alec Pestov
to track financial performance of your project so that you know whether you're losing money or you're making money on your project at the end of each day.
9:14
Andrea Ciaffi
Interesting. And now we are on a podcast. It's a bit hard to describe what a mobile
9:20
Alec Pestov
I'll send you the videos. I'll send you the videos.
9:20
Andrea Ciaffi
app or piece of software looks like. Yeah, but we'll put the link to the videos. Can you try to explain to us, like, I'm trying to imagine the life of somebody using VSite. They're on a construction site, they're on a site, the construction site, somewhere outside.
9:38
Andrea Ciaffi
What does it look like and where does the value show up first for them? Where do they, when do they have these aha moments where they realize, okay, this is actually doing something good for me?
9:51
Alec Pestov
We have multiple. features even within the mobile app. So if you look at the first step, which is packaging data and delivering it to the field, our system can organize data from multiple sources like municipal GIS databases, BIM data, isolated cut files, PDFs.
10:08
Alec Pestov
So all of that data. goes into one system and produces digital twin that is constantly updated in the cloud. So it's not a static digital twin, but it's an aggregation of different documents and file formats that you can access in many ways. And we have two types of users.
10:27
Alec Pestov
We call them power users and regular users. Now let me start with regular users. A regular user is somebody who spent his or her entire life with the paper-based drawings, walking around and looking at the drawings and measuring with the measuring wheel or measuring tape.
10:47
Alec Pestov
Many people are very effective at building that way. And there is no need for us to re-educate the entire industry and change millions of people. If they're used to that, we target that audience first. So what we do is, except we take the paper-based PDFs they typically have, and from paper format, it becomes a digital PDF.
11:09
Alec Pestov
The advantage to them is that we have a bunch of features that allow them to see themselves working on that PDF. So whether if you're looking at the paper-based sheet plan on construction site, you need to figure out where you are in relation to all the objects there.
11:26
Alec Pestov
In our case, you standing physically, you see your arrow, and when you walk, you see yourself walking on that PDF. Again, it's not a huge departure from holding paper in your hands, except you're looking on your phone or your tablet, but it's the step in the digitization, in the digital direction.
11:42
Alec Pestov
And it makes it easier because we have several features that will make it better. For instance, we have PDF comparison, where the system can produce, isolate only the changes from the previous version to the current version of the PDF. So, when you get the new package sheets of a version set from your designer.
12:03
Alec Pestov
instead of going through everything, understanding everything, you know that this manhole moved from here to here. And then because you're looking on the phone and you can see yourself on that map, you can physically perform clash detection in the field. So you say, oh, it was originally here, but now it moved here, and now we have a tree, and tree requires six months of permitting.
12:23
Alec Pestov
and the $1, 000 for permits changes the cost calculations and everything else for the project. So that's already a step forward. And it doesn't require a lot of explanation because it's the same thing that you look at on paper, except it's digital and has a lot more things that you can utilize directly from the system.
12:37
Alec Pestov
So this is what we offer to Regular users. To power users, you can switch from the same device. Just click one button, you switch into AR, AR mode, and you can see what you have around you and what you need to build through a centimeter accurate AR.
12:59
Alec Pestov
And that is very unique to us because most solutions, AR solutions, they're based around QR codes or a really simplistic integration with high accuracy GPS devices. In our case, you actually see visuals with centimeter level accuracy. We have patented mechanisms for aligning AR with the world so that it's perfectly aligned, which is a fairly difficult thing to accomplish.
13:26
Alec Pestov
And then you can leverage this to build better. Now, how it's used in the field? We have a recent example from a data center construction. So it's actually on our website in the new section. The company needs to build utilities leading, going into the data center.
13:46
Alec Pestov
and they have to build it within 10 centimeter tolerances. And when you're working off paper sheets, which is common, and you're standing in the big open field, it's very difficult to get all these tolerances in. Of course, you can have control points, but if you have discrepancies between the control points.
14:08
Alec Pestov
and what you need and what the control points should be and they had on that side specifically, then what do you trust? How do you even validate that you have things built accurately? And because in our tool, you can see what you need to build, you can stake out from that model.
14:25
Alec Pestov
then it makes it very easy to see even the discrepancy of two, 3 centimeters. And that immediately gives the general contractor idea of where they might be doing something wrong. So this is kind of the second thing, is you look at your tablet, I'll send you the videos, and you see the world behind it, just like when you take a picture or video with your phone.
14:51
Alec Pestov
except you see augmented reality superimposed on that with all the details so you know exactly where different conduits, pipes, manholes supposed to be.
15:01
Andrea Ciaffi
Yeah, very cool technology. And many people may be familiar with similar technologies, but inside buildings where people use a tablet to overlay services, but you actually deliberately target outdoor situations. And can you share with us where do you exactly draw the line? Where do you actually stop?
15:23
Alec Pestov
The problem with indoor AR is positioning. You want to localize yourself with high accuracy. You want to know where you are in relation to the model. And when you go inside the building, there is no single source of truth that you can use to guarantee your alignment.
15:42
Alec Pestov
Yes, you can use walls, for example, and align against the walls. But what if walls shifted and the model doesn't reflect it? Like especially dry walls, it can be, you know, Half a meter this way, half a meter that way, it's fluid. Now, you can use QR codes, which some companies use.
16:04
Alec Pestov
The problem is the moment you scan the QR code and start moving, you will start experiencing what's called AR drift. which is your device miscalculates your movements in space slightly.
16:21
Alec Pestov
But even if you have miscalculation of 2%, so for example, you traveled 10 meters and the system thinks that you traveled 9 meters and 80 centimeters, so 2% mistake, but you're still 20 centimeters off at the end of the 10 meter walk. And this is an inherent problem with AR indoors, is that there's no source of truth.
16:43
Alec Pestov
You're relying on your starting point, and then you don't know if you drifted away or not. So that's number one. Number 2 is tolerances. Indoor, your tolerances are Different than outdoor when you're installing, let's say, a meter-wide sewer pipe. If it's 5 centimeters to the left, 10 centimeters to the right, doesn't make a huge difference.
17:10
Alec Pestov
As long as the slope is fine, as long as it's there, it's fine, it's within the tolerances. Inside the building, you may have one wire and next wire within like centimeter, couple of centimeters.
17:24
Alec Pestov
And if you have the tolerances that AR cannot guarantee at the moment, say, if you have the centimeter level tolerances and you cannot guarantee positioning and you cannot guarantee accurate placement, then you may give false confidence to the end user and they say, oh, I know exactly where the point between 2 wires, they drive nail and they hit one of the wires.
17:45
Alec Pestov
So we don't want that to happen. So the positioning issue, and much tighter tolerances indoors, make it. non-practical in my view for indoor use and that's why we deliberately don't go indoors.
18:01
Andrea Ciaffi
Do you rely at all on GPS? Because I'm conscious that GPS doesn't really work inside buildings. But do you use it in your application?
18:08
Alec Pestov
Yes. Yeah, so we use different types of GPS. We say have, we support 40 or so different receivers. Our solution is very simple. So you take your phone, take GPS, Snap one or not, now I have a centimeter accurate form.
18:25
Andrea Ciaffi
Nice.
18:26
Alec Pestov
And it makes it easy for me because I can put it in my surveying vest and it can be hands-free to build whatever I need to build. When I need to validate, I just pull it out, go like this. Now I have some to me accurate phone.
18:39
Alec Pestov
The same tool I can use with the surveying pole, so I just screw it to the bottom. And now I'm ready for redlining or staking out. So I have one universal solution that I can carry in my vest, including pole, because they can be collapsible.
18:53
Alec Pestov
So you can put it in your back pocket. So, and GPS obviously needs a clear view of the sky, otherwise it doesn't work. So you cannot go indoor with that. So that's why we stay outdoors.
19:08
Andrea Ciaffi
Like, very, very nice. Like, it's a real tangible, real world application of real technology, very cool stuff. People may be familiar with services like that before digging in Australia, for example, or 8 by one that most retail construction teams where the utilities are.
19:23
Andrea Ciaffi
How do you explain? I would explain what this side does that those services don't.
19:33
Alec Pestov
Depending on your country and the region, most of the world has similar services. So you have dialed before you dig. In Canada, we have called before you dig. In the US, they have 811. In the Netherlands, in the Netherlands, they have click system.
19:47
Alec Pestov
So there are different systems around the world. but conceptually. other than the click. So click is unique in that space. Most services dial before you degrade 1 once. They're really a dispatcher. What they know is they know, they take your request, incoming request, they will write it down, and they know which utilities operate in the area of the request.
20:16
Alec Pestov
And they will play, and they will place a request to the utilities to dispatch a locator to mark utilities on the ground. So either internal locators or third-party locators will come in on behalf of utility and mark utilities on the ground. So in that space, dial before you dig, call before you dig, 811, they act as a married dispatcher.
20:37
Alec Pestov
So this person call, you guys go do the work. They don't own any data. They cannot provide any data in most cases. They don't have a centralized repository of all utilities. They just know who operates in that area and who to call. To submit locators.
20:56
Alec Pestov
The only difference in that is click. So is the K-L-I-C. It's the system in the Netherlands, and it's a government registry that has all utilities for the entire country. Click operates differently because when you place a request to click, they actually deliver data to you in a readable format, so you can use your data yourself.
21:19
Alec Pestov
How you use it, you decide. But you know, click distinction is that dial before you dig is just a dispatcher. Now, how we differ from that is that we don't operate in that space. We don't know who to call to for utility locates.
21:34
Alec Pestov
We don't provide utility locate service ourselves. So we don't really directly compete with them. But in, we piggyback on the work they do, because when they dispatch locators. Locators mark the ground, so they leave markings on the ground. And then the excavator comes in, looks at the markings, looks at the report, and starts digging, trying to avoid utilities there.
22:02
Alec Pestov
Quite often on the... especially brownfield type of projects, you will have HVACing, so somebody will do, and the terms vary from region to region, it's going to be potholing, daylighting, test trenching, whatever the term you may use, but essentially will bring HVAC, HydroVAC, and they will excavate.
22:25
Alec Pestov
Call to validate the security is there, and they will document it. Now, what we offer into that space is that our solution, which runs on phone or tablet, you can go ahead and scan the utility, scan the locates, sorry, scan the locates. And when you hydrovact and you expose the utility to validate it, you can scan it too.
22:48
Alec Pestov
And that all goes, so all you have to do is it's a single button operation, so it's easier than to use your phone camera. So it's the same concept, press one button and it's done. But then you have a digital record of what was on the ground and how the trench looked like and what utilities looked inside that trench.
23:07
Alec Pestov
Now you preserve that knowledge for situations when you have to review it again. So let's say your crew's rotated. Now we have a new foreman, new crew on site. They don't need to redo the process. They can rely on the records from previous to compare them against locates done now.
23:22
Alec Pestov
And if somebody did a mistake, made a mistake, and they incorrectly mark the ground, you can pull up the records and say, hold on, these two don't match. and then it may trigger additional investigation for you. So what we do, we don't compete with Dial Before You Dig, to go to your question originally.
23:43
Alec Pestov
They operate completely independently in isolation from us, but we leverage output they produce to help companies avoid damages.
23:52
Andrea Ciaffi
Yeah, and for a second, let's think about a world where your platform, your tool doesn't exist, right? So what is the risk today when data leaves only on the ground as a spray paint and nowhere else? What can possibly go wrong?
24:09
Alec Pestov
Oh, so many things, so many things. First of all, legal liability. Who is responsible for damages if something happens? It is not uncommon to go through an investigation process or arbitration process when there was a major damage, and then you have to argue whether the paint was on the ground, whether it was in the right place, because there are...
24:33
Alec Pestov
different liability levels and different liability guidelines depending on how it was painted, where it was painted, whether you read the markings correctly. So that's one.
24:44
Alec Pestov
In our case, even if you do it at the locate level, and you scan it, you have evidence, concrete evidence that on such and such date, such and such location, because it's all geo-referenced with centimeter level accuracy, you had paint on the ground here and it's something you cannot argue with.
25:06
Alec Pestov
Another issue, it's almost like an insurance policy for the locator. We had quite a few situations when locators will do the job correctly and then excavator will hit the utility and they will step out and spray paint new line and say, yeah, you see it was right here.
25:22
Alec Pestov
So I was working off this way. Humans are humans, situations like that happen. Our system offers essentially an insurance policy for Locator that, yeah, I documented it this way. And also on the GC side, the benefit is for crews to preserve location.
25:44
Alec Pestov
There are many actually benefits, not just for one crew, because the GC, they can continue utilizing the same data set for the duration of the project, minimizing time required for different crews to get up to speed, minimizing clash detection times and everything else.
26:02
Alec Pestov
In other benefit is that they can sell the data or provide it as part of the deliverable to the owner. And then owner, for engineering reasons, purposes may know where utilities are. So next time they're excavating on the same St. or adjacent St. they're not starting from scratch, but they have historical data.
26:23
Alec Pestov
We shortcut some of the engineering efforts and the original and initial investigations. So there are many benefits of documented across the entire value chain. And if our tool doesn't exist, well, it doesn't exist in most places. So you just do what you do.
26:34
Alec Pestov
You call, dial before you dig every single time, and you put fresh paint on the ground.
26:43
Andrea Ciaffi
Hopefully, you can expand your footprint soon and solve the problem for many people in different parts of the world. And speaking of that, you have created a pretty sophisticated piece of tech, right? And you're trying to give this tool to an industry that is familiar with more simple tools like hammers or like Very simple stuff, and they tend to be resistant to change often.
27:07
Andrea Ciaffi
So, do you think that this concept has shaped the way you design a soft like piece of software for people like that?
27:20
Alec Pestov
I actually posted a couple of videos on LinkedIn about that. One thing that companies that design software for construction industry cannot afford. is not spend time physically in the field with people who use it. We catch ourselves making. Right now we make fewer mistakes, but we catch ourselves making mistakes in the system design.
27:47
Alec Pestov
Because what looks good on the big bright screen in the laboratory or UI designer works completely different in the field. And what happens in the field is that Office workers, they have time and tolerance for mistakes. So if you have a browser that doesn't work, you will just shut it down and restart.
28:09
Alec Pestov
If you have a computer that has a glitch, you just reboot and wait for that. And yes, it's annoying, but you will reboot and you will reboot again and you will reboot again. In construction, there is a lot of pressure to perform. You have to deliver.
28:26
Alec Pestov
You have no time to waste. And you have tools that have been working historically, reliably. So what I find is that it's not so much a hammer and something else, but if you have a tool that works, yes, it may take longer and more effort to do this.
28:42
Alec Pestov
And if somebody offers you a better tool, supposedly, but it doesn't work, immediately, then in construction people will set it aside and will start doing what they did historically. They have no time to investigate this if it didn't work. It makes it very difficult.
28:57
Alec Pestov
for us because we have to spend extra time working with everybody and fixing all the bugs. So more so than if we would have tailored it to the office workers. But there, so this crowd is very demanding. You have to be really closely aligned with your customer.
29:17
Alec Pestov
watch them use your software yourself. So sometimes you just give it to them and observe and record what they do. But that's the only way you can overcome that resistance to change. If you have better tool and it works, rational people will jump on board.
29:34
Andrea Ciaffi
Yeah, and speaking about augmented reality, AR, for a long time it's been carrying a lot of hype. What do you think, what can construction grade AR generally do on a site today that like a just simple tablet just simple simply full of PDF cannot do today?
29:52
Andrea Ciaffi
Because you also mentioned like you also have a part of the product that is probably based on PDF, but you also have a more sophisticated version where. It's related to just using AR. What's the main difference between these two and what can you achieve with AR that you cannot achieve with simple PDF?
30:14
Alec Pestov
AR gives you quite a few additional tools that PDF can't. To give you an example, grading. in the road construction. You have your grade, you have existing grade. How do you know whether you need to cut or you need to fill? like quickly.
30:31
Alec Pestov
Of course, you can do all the calculations and measurements. That's the way it's done, and it will be done. But quickly, if you want to figure out where you have to cut, where you have to fill. With PDF, you can't. You have, of course, your elevations and grades mapped on your PDF.
30:48
Alec Pestov
So you can read the plan, you can call surveyor, you can work with total station, measure everything. and then get the answer. But it will take you hours, if not days, and a lot of effort. With AR, and I can send you a couple of videos of how it looks, we can display the grade as it's supposed to be.
31:10
Alec Pestov
And we have a tool built right into the system. So it's just as easy as sliding a slider on the screen that scans the world around you. And then you can instantly see if your proposed grade and the actual ground level, if they match, if one is higher than another, and you can see the difference between them.
31:29
Alec Pestov
And that instantly gives you understanding of whether you have to cut or fill in that location, and you take a few steps forward and do the same thing. So you can see how your grade relates. So it's one of the examples. Another example is understanding complex 3D elements.
31:45
Alec Pestov
Infrastructure sometimes overlaps. at different levels, and you want to understand how much clearance you have. And our system allows you to clearly see where things are. And the last thing is just the general guidance.
31:58
Alec Pestov
If I can see clearly that I'm standing on top of the pipe and it goes there and then it curves this way, and I can follow that, it's a lot easier than standing with the PDF.
32:09
Alec Pestov
And then measuring from the control points with the wheel, drawing things on the ground, takes a lot longer and it doesn't produce that visual memory of where things are at this particular location. Because if you see it, you remember it better. But if you measure it, Just measurements, those are things that are harder to retain in your memory.
32:32
Andrea Ciaffi
Yeah, and... From this conversation and from your past history that you shared with us, I can see that you are a tech enthusiast, somebody who really likes technology. So I really need to ask you questions about AI. And I just wonder like, is there any real application where you're using AI inside this site or?
32:53
Andrea Ciaffi
could be anything related to how your team is improving their productivity or could be AI in product.
33:01
Alec Pestov
We do utilize AI and we will release quite a few AI components. The trick about AI is that when... Generally, you think about AI, you think.. . ChatGPT or Claude, Gemini. So those are the options. You think of the screen where you type your questions and then answer start listing.
33:24
Alec Pestov
So that's a general understanding of AI. Our AI works, or not just our AI, but AI can work in different ways. So, for example, we use AI to recognize packing slips. You receive pack and slip for each delivery. Now you can of course take the pack and slip and type it into Excel yourself, which is what most companies do.
33:47
Alec Pestov
And a typical site will spend two to three hours a day typing pack and slips into Excel every day. In our case, you take a picture and we have an AI-driven workflow. that digitizes entire packet sleep and creates that Excel entry, well digital entry, and matches it to the course code.
34:04
Alec Pestov
So. a lot easier. Now, you don't know it's AI, you just see a magic of taking a picture and see it's cloaking on the background and say, oh, here's the full digital copy of that. But it's AI behind the scenes. Another good example is we worked in Germany with telco companies.
34:27
Alec Pestov
And in Germany, there are strict regulations around... trench profiles for laying fiber cables. It has to be no deeper than, it has to be no shallower than, you have to have specific depths to cover.
34:42
Alec Pestov
So if you measure from the top of the cable to the end of the trench, it has to be specific depths, it has to be specific width, it has to be in the right location. If you have an embankment on the side, then you have to have so many centimeters between that and the trench so that you can safely walk.
35:01
Alec Pestov
So all of that is part of the trench requirement. And the way... German inspectors for telcos used to handle projects is they physically would go to every site every day, and they will measure with the measuring tape, all those parameters, write them down, and then the trench will be approved.
35:19
Alec Pestov
Now, of course, it's not practical because Germany is fairly large country territorially. And you may have one site here, another site 3 hour drive. So how many sites can you handle? Only two sites as an inspector. So you have one inspector for two sites and everybody is waiting for the inspector to arrive.
35:39
Alec Pestov
And in the past, when situations happened and companies were allowed to backfill before inspector arrives, They had numerous instances where the same trench had to be reopened twice, sometimes three times, to collect acquired evidence before your invoice will be paid out. Now, there's an efficiency everywhere because you're waiting, you're measuring manually.
36:02
Alec Pestov
And if you're relying on pictures, then one of the reasons why telcos could not rely on pictures is that dishonest contractors will take a picture in one place, take a picture, then skip the bet. part of the trench, take another picture, and everything looks good in pictures.
36:24
Alec Pestov
But there was a problem that was missed in the middle. Now, we developed an AI component that analyzes the entire trench profile and to see if it's within the parameters. And what it does, it benefits both parties, the inspector and the owner and the GC.
36:39
Alec Pestov
For GC, they scan it with our tool. About 40 seconds for every 30 meters, there's some processing. So 2 minutes later, you know the results. And it highlights the areas within the trench. where you have potential problems. And then you can go and take additional pictures or measurements there and submit it.
37:02
Alec Pestov
Now, from the owner perspective, they receive a geo-referenced 3D scan of the trench with problems highlighted and immediately with pictures attached proving that those are not problems. Now, you don't have to dispatch this inspector to every location because you can't really fake that scan.
37:21
Alec Pestov
Scan is geo-referenced, timestamped. So you can just keep a picture of the trench on your phone and submit it as a proof. It was taken here and there. And it provides instant analysis to both parties for them to make a decision and saves you from reopening trenches, saves you from driving to each site every day.
37:40
Alec Pestov
So there are huge benefits. Now, do they know it's AI? No, they don't. So we don't advertise it as AI. All they know is that they press the button, they scan it, say upload, and a minute later they receive results and say, hey, go here and check it again.
37:52
Alec Pestov
But it's AI in the background. So AI is a really powerful tool. You just need to use it. correctly for construction purposes.
38:03
Andrea Ciaffi
That's great. And speaking about AR, augmented reality, this technology has been around for a few years now. And where do you see AR on the app cycle right now?
38:19
Alec Pestov
Well, the same place as, I would say, NFTs or... Remember in a few years ago?
38:26
Andrea Ciaffi
Yeah.
38:28
Alec Pestov
So it's about in the same place. So it's very unpopular. It's there, it exists, but it's probably at the bottom of the hype cycle. And now we probably start exiting. The thing is, pure AR has usability in many instances.
38:44
Alec Pestov
When you buy, for instance, glasses, sunglasses, it's a lot easier for you to have an AR based fitting mechanism where you can see yourself wearing those glasses in real time. So you can quickly narrow down to the glasses you like. It's practical. So AR like that exists and will continue expanding.
39:03
Alec Pestov
On the industrial side, AR. has been used, but certainly not at scale. So I would say in terms of the hopes that everybody had versus where it is now, we're absolutely at the bottom of that hype cycle.
39:19
Andrea Ciaffi
And it's very hard to predict the future, but where do you see AR in the next 5 to 10 years?
39:28
Alec Pestov
The main limitation for AR right now is equipment. As you can see, to get AR, I need to carry 2 devices. If I had just the phone and it was as easy as turning on my phone and then seeing AR, then it will be a lot more widespread.
39:43
Alec Pestov
So I think once technology catches up so that you can start eliminating the time it takes to launch AR, you will see a lot more uses for AR. So I think in five years, I don't think it's going to conquer the world. So certainly it doesn't develop at the pace of AI, but I think it will become more common in quite a few areas.
40:05
Alec Pestov
So you won't necessarily see it. You will see it as part of your existing toolkit, but it will become much more widespread and common in some areas.
40:16
Andrea Ciaffi
That's good. Alec, we are at the end of this episode. If people want to reach out to you, to VJS, where can they find you?
40:24
Alec Pestov
Please go to our website, www..vgis.io, or you can send me an e-mail, or I'm on LinkedIn, very active on LinkedIn. Feel free to DM me on LinkedIn. I will be happy to chat.
40:39
Andrea Ciaffi
Alec, thanks for joining us on the Digital Construction Podcast.
40:43
Alec Pestov
Thank you for having me.

15 April 2025
Online Training and Digital Adoption with Susan Brattberg
Understand why a comprehensive training plan is a crucial insurance policy for technology investments